Market Timing: Transaction Costs and Taxes

I continue to look for criticisms of market timing and see how the market timing proponents would answer them. Last time I looked at some criticisms from Vanguard founder John Bogle. I think those criticisms were answered reasonably well.

Another major area of criticisms centers around frequent trading, transaction costs and taxes. Critics say market timing creates too many trades. Too many trades incur transaction costs. If those trades are done in a taxable account, profitable trades also produce capital gains, which are taxed. Any extra profits over buy-and-hold will be overwhelmed by the transaction costs and taxes.

This is true for some forms of market timing, if they have you trade in and out frequently in short successions. It’s also true in the past when transaction costs were high and when access to tax deferred accounts were limited. It’s not necessarily true for all market timing or as relevant when transaction costs have come down a lot (sometimes zero) and when investors have most of their money in tax deferred or tax free accounts.

Take the “sell in May and go away” market timing strategy for example. It says instead of holding the same portfolio all year long, one should sell stocks in May and buy back in November. That’s two trades per year per position. If one uses two broadly diversified ETFs for stocks (one US, one international) and pays $8 per trade, that’s $32 a year in commission, plus maybe 0.05% in bid/ask spread. The transaction costs are very low (~0.10% a year for a $50k portfolio).

Transaction costs are zero if one uses open-end index funds, because there won’t be any commission or bid/ask spread. Even the buy-and-hold stronghold Vanguard will allow two trades per year per fund. Its frequent trading policy only restricts buying back within 60 days. Buying back after six months won’t be a problem.

When investing meant buying individual stocks instead of broadly diversified index funds or ETFs, market timing was expensive. Selling a 100-stock portfolio meant paying 100 trading commissions. With broadly diversified index funds that invest in thousands of stocks in one shot, transaction costs are no longer an issue.

When people had defined benefit pension plans, investing pretty much meant investing in taxable accounts. There were IRAs since 1970s but the contribution limit was only $2,000 a year until 2001. Taxes make market timing in taxable accounts very expensive. People have defined contribution plans now. IRA contribution limits have also more than doubled. As a result, people have most of their money in tax sheltered accounts. If one makes market timing trades in a 401k plan or IRA account, there won’t be any taxes.

I wonder if the transaction costs and taxes arguments against market timing are a carryover from a previous era when transaction costs were high and taxes were unavoidable. I looked up a few academic studies on market timing. The assumed transaction costs are typically 0.5% per one-way trade. Two trades per position per year would cost 1.0%, instead of under 0.10% or zero for index fund investors at this time.

What makes or breaks a market timing strategy like “sell in May and go away” won’t be transaction costs or taxes, because both are avoidable. Whether it will be a successful strategy depends on whether the historical pattern will hold in the future.

We can see although it doesn’t work every year, on balance it worked in the last 40 years or so. Will it work in the next 40 years? Nobody knows. How can anyone give any evidence that it won’t work when the future hasn’t happened yet?

Will I use “sell in May and go away”? Probably not, but it won’t be because of transaction costs or taxes.